JP2016099949A - Meta-information writing method, meta-information reading method, file management system, computer system, program, and data structure - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a meta-information writing method for storing meta-information in a storage medium, a meta-information reading method, a file management system, a computer system, and a program.SOLUTION: A meta-information writing method is a method for writing meta-information in a storage medium in which an index that retains meta-information of each of one or more data items is preserved. The meta-information writing method includes step S101 at which a computer system 110 responds to a meta-information writing request, and decides whether a capacity for meta-information relating to the writing request meets a criterion for preservation outside an index, step S103 at which if a decision is made that the criterion for preservation outside an index is met, at least part of the body of the meta-information relating to the writing request is written in a storage area outside the index, and step S104 at which position information representing the storage area in which at least part of the body is written is stored in the index.SELECTED DRAWING: Figure 5

Description

  More particularly, the present invention relates to a meta information writing method, meta information reading method, file management system, computer system, and program for storing meta information in a storage medium. .

  A format and software called LTFS (Linear Tape File System) that enables access to a tape medium through a general file system interface is known. The LTFS format is an open specification, and the implementation of a file system that supports tape media written in a format according to the LTFS format is an international business machines corporation (International Business Machines Corporation, IBM (registered trademark)). , Oracle and Hewlett-Packard. By using LTFS, it is possible to access a tape medium from an application that has previously used a hard disk drive or flash memory without any change.

  In the LTFS format, a tape medium is used by being divided into two partitions called an index partition and a data partition (for example, JP 2012-65087 A (Patent Document 1)). The latest index is written in the index partition, and the main data of the file and the history of the index so far are written in the data partition. The index holds file attribute information such as a pointer to the main data on the data partition and a file name. Without the index, the main data of the data partition cannot be accessed, so the index is written on the tape medium regularly or irregularly.

  It is known that the capacity of the index is generally about 1 kB per file (that is, about 1 GB for 1 million files). However, this is a typical example and may vary depending on the actual capacity of each meta information. For example, a fork such as ADS (Alternate Data Stream) or resource fork can have an arbitrary capacity, and thus may have a capacity comparable to the file body. When a file having such meta information as fork is used, the size of the index may increase.

  For this reason, as the size of the index increases, it takes time to write the index itself, and it may happen that the index does not fit in the index partition of a predetermined size. Further, as described above, since the index is cumulatively written in the data partition, the resource of the tape medium is compressed due to the increase in the size of the index. Accordingly, it has been desired to develop a technique capable of devising storage of the large-capacity meta information as described above so as not to press the resources of the tape medium.

JP 2012-65087 A

  The present invention has been made in view of the insufficiency in the prior art described above. The present invention increases the size of an index in a storage medium storing an index that holds meta information of each of one or more pieces of data. An object of the present invention is to provide a meta information writing method, a meta information reading method, a file management system, a computer system, a program, and a data structure capable of preventing and storing meta information without squeezing resources.

  In view of the above problems, the present invention provides a meta information writing method having the following characteristics. This meta information writing method is a method for writing meta information to a storage medium in which an index for holding meta information of each of one or more pieces of data is stored. In this meta information writing method, in response to a meta information write request, the computer system determines whether or not the capacity of the meta information related to the write request satisfies the criteria for storage outside the index. If the computer system determines that the non-index storage criterion is satisfied, the computer system writes at least a part of the meta information main body related to the write request to a storage area outside the index, and at least a part of the main body. The position information indicating the storage area in which is written is stored in the index.

  The present invention also provides a meta information reading method for reading meta information from a storage medium in which an index is stored. In this meta information reading method, the computer system determines whether the meta information related to the read request is stored outside the index in response to the meta information read request. When it is determined that the computer system is stored outside the index, the computer system acquires position information indicating a storage area outside the index from the index, and the storage area specified based on the acquired position information Then, at least a part of the main body of the meta information related to the read request is read.

  Furthermore, according to the present invention, there is provided a file management system that performs file management of a storage medium in which an index that holds meta information of each of one or more pieces of data is stored on a computer system. In response to the meta information write request, the file management system determines whether the capacity of the meta information related to the write request satisfies the non-index storage standard, and the non-index storage standard. A write processing unit that writes at least a part of a meta information main body related to a write request to a storage area outside the index when it is determined to satisfy, and a storage area in which at least a part of the main body is written And a storage processing unit that stores position information in the index.

  Further, according to the present invention, the file management system, in response to the meta information read request, determines whether the meta information related to the read request is stored outside the index, and the index When it is determined that the data is stored outside, an acquisition processing unit that acquires position information indicating a storage area outside the index from the index, and a read request from the storage area specified based on the acquired position information And a read processing unit for reading at least a part of the meta information main body.

  Furthermore, according to the present invention, an interface device for connecting to a processor, a memory that communicates with the processor, and a storage medium drive that drives a storage medium in which an index that holds meta information of each of the one or more data is stored A computer system is provided. In response to the meta information write request, the computer system determines whether the capacity of the meta information related to the write request satisfies the criteria for off-index storage, and satisfies the criteria for off-index storage. If the determination is made, at least a part of the main body of the meta information related to the write request is written in the storage area outside the index, and the position information indicating the storage area in which at least a part of the main body is written is stored in the index. Composed.

  Further, according to the present invention, in response to the meta information read request, the computer system determines whether the meta information related to the read request is stored outside the index, and stores the data outside the index. If it is determined that the storage area is determined, the position information indicating the storage area outside the index is obtained from the index, and the body of the meta information related to the read request is obtained from the storage area specified based on the obtained position information. Is configured to read at least a portion of.

  Furthermore, according to the present invention, it is possible to provide a program for causing a computer system to function as each part of the file management system.

  Furthermore, according to the present invention, an index partition that stores an index in which one or more meta information about each of one or more files is stored, main data of each of the one or more files is stored, and an index history is added. A main partition of meta information having a capacity satisfying a predetermined standard, and an index having a capacity satisfying the predetermined standard. It is possible to provide a data structure characterized in that it retains position information indicating a storage position on the data partition in which is written.

  With the configuration described above, in a storage medium storing an index that holds meta information of each of one or more pieces of data, the index can be prevented from being enlarged and the meta information can be stored without squeezing resources.

  Other effects will be understood from the detailed description of the invention below.

The figure which shows the volume layout of the tape medium used with the file system for tapes of a prior art. 1 is a system configuration diagram of a tape file management system according to an embodiment of the present invention. FIG. The hardware block diagram of the computer system which comprises the file management system for tapes by embodiment of this invention. The software block diagram of the file management system for tapes by embodiment of this invention. 6 is a flowchart showing a meta information writing method executed by the computer system according to the embodiment of the present invention. The figure which illustrates the data structure of the index file created based on the meta-information writing method shown in FIG. 5 is a flowchart showing a meta information reading method executed by the computer system according to the embodiment of the present invention. The figure explaining the structure by which the index management part by embodiment of this invention maintains index data on memory. The figure explaining the effect of the meta information storage system which the file management system for tapes by an embodiment of the present invention provides with a data structure.

  Hereinafter, although embodiment of this invention is described, embodiment of this invention is not limited to embodiment described below. In the embodiments described below, the meta information writing method, the meta information reading method, and the meta information writing method executed by the tape file management system that performs file management of the tape medium, respectively, as the file management system A meta information reading method and a tape file management system will be described as an example.

  FIG. 1 is a diagram showing a volume layout of a tape medium used in a conventional tape file system. FIG. 1 shows a layout based on the LTFS format supported by LTO5 or later as an example. In the LTFS file system, it is necessary to initialize the tape medium in the LTFS format in advance when reading and writing files on the tape medium. The tape medium is logically divided into two types of partitions called an index partition (hereinafter referred to as IP) and a data partition (hereinafter referred to as DP) by initialization as shown in FIG. Divided and used.

  In the volume layout shown in FIG. 1, “FID” records the same information as that written in the cartridge memory of the tape cartridge, which is a tape medium. FM represents a file mark written to indicate a file delimiter. The “VOL1 Label” record stores information such as a label identifier, a label number, and a volume identifier. The “LTFS Label” record is an XML (eXtensible Markup Language) data structure that stores information on an LTFS volume and an LTFS partition.

  FIG. 1 shows a state in which two files (“File 1” and “File 2”) are written on the tape medium. The latest index (“Index # 2” in FIG. 1) is written in the IP, and the file main data (“File1” and “File2” in FIG. 1) and the history of the index (FIG. 1) are written in the DP. "Index # 0" to "Index # 2") are written therein.

  In the LTFS format, the directory structure of the file, the logical position information where the file is written on the tape, the file name and other indexes, and the index that holds the meta information of the directory and symbolic link are held as an index file in XML format. ing.

  In the LTFS format, if there is no file meta information (logical position information) in the index file, it cannot be determined where the file body data exists. That is, even if file data is written, if the index file is not written to the tape medium, the file cannot be read from the tape medium if the system is shut down as it is. Therefore, in the LTFS, from the viewpoint of protecting the written file, there is a setting in which an index file is written to the tape medium every predetermined time or an index file is written to the tape medium every time the file is closed.

  On the other hand, if the index is written too frequently, it will squeeze the resources of the tape medium. It is known that the capacity of the index is generally about 1 kB per file, for example, about 1 GB for 1 million files. At this level, even if the index is frequently written, it does not become a big problem. However, especially when using files with meta information that has a capacity comparable to the file size, such as a fork (ADS or resource fork), and in some cases larger than the main data, the increase in index size is as follows. It may lead to the defects shown. In other words, an increase in the size of the index may cause a longer time for index writing itself and may cause a situation in which the index does not fit in the IP. Increasing the size of the index can also cause cumulative resource pressure due to the cumulative storage of the index in the DP.

  Therefore, in the tape file management system 100 according to the embodiment of the present invention, in response to a write request for predetermined meta information of predetermined data (meta information can be attached to a file, a directory, and a symbolic link, respectively). A configuration is adopted in which it is determined whether or not the capacity of the meta information related to the write request satisfies the criteria for storage outside the index. If it is determined that the meta information is a fork or the like described above having a large capacity and satisfying the criteria for storage outside the index, the main data of the meta information related to the write request is stored on the DP outside the index. The logical position information indicating the storage area on the DP in which the main data of the meta information related to the write request is written is stored in the index. This prevents index enlargement, avoids resource pressure caused by the above-described cumulative storage of indexes, and efficiently stores meta information.

  The meta information management function realized in the tape file management system 100 according to the embodiment of the present invention will be described below with reference to FIGS.

  FIG. 2 is a diagram showing a system configuration of the tape file management system 100 according to the embodiment of the present invention. FIG. 2A illustrates a tape file management system 100A using a single tape drive. In contrast, FIG. 2B illustrates a tape file management system 100B using a tape library including one or more tape drives.

  A tape file management system 100A shown in FIG. 2A includes a computer system 110A and a single tape drive 132A. A tape medium 134A such as a tape cartridge is inserted into the tape drive 132A. The computer system 110A accesses the inserted tape medium 134A via the tape drive 132A. The tape drive 132A constitutes the storage medium driving device in the embodiment shown in FIG.

  The tape file management system 100B shown in FIG. 2B includes a computer system 110B and a tape library 130B. The tape library 130B includes one or a plurality of tape drives 132B-1 to 132B-L, and manages one or more tape media 134B-1 to 134B-O taken in the library. The robot mechanism selects a desired tape medium from one or more tape media 134B in the library and inserts it into the tape drive 132B. The computer system 110B accesses the tape medium 134B via the tape drive 132B. The tape drive 132B constitutes the storage medium driving device in the embodiment shown in FIG.

  2 (A) and 2 (B) show the tape file management system 100 in two modes. However, in any of the modes and other modes, details of the present invention will be described later. A meta information writing function and a meta information reading function according to the form can be implemented. The system configuration shown in FIG. 2A corresponds to the LTFS software provided by IBM (registered trademark) as LTFS SDE (LTFS Single Drive Edition). The system configuration shown in FIG. 2B corresponds to the LTFS software provided by IBM (registered trademark) as LTFS LE (LTFS Library Edition).

  Although detailed description is omitted, distributed sharing such as GPFS is performed on a computer system including one or more computers such as provided by IBM (registered trademark) as LTFS EE (Enterprise Edition). The LTFS software that cooperates with the file system can also implement the meta information writing function and the meta information reading function according to the embodiment of the present invention.

  FIG. 3 is a diagram showing a hardware configuration of the computer system 110 constituting the tape file management system 100 according to the embodiment of the present invention. As shown in FIG. 3, a computer system 110 includes a CPU (Central Processing Unit) 10 that is a processor, a memory 12 such as an SDRAM (Synchronous Dynamic Random Access Memory), a ROM (Read Only Memory), and an HDD (Hard Disk). Drive) and a flash memory, a network interface device 16, and an input / output interface device 18.

  The network interface device 16 is configured as a network adapter such as a NIC (Network Interface Card), and connects the computer system 110 to a network such as a LAN (Local Area Network). The input / output interface device 18 is configured as a network adapter such as NIC or an HBA (host bus adapter) such as FC, and connects the computer system 110 to the tape drive 132 or the tape library 130.

  The computer system 110 implements functions and processes described later by reading various programs stored in the storage device 14 and developing them in the memory 12 under the control of the CPU 10. The hardware configuration of the tape drive 132 and the tape library 130 is not particularly shown, but can be configured with a CPU, memory, ROM, network interface, input / output interface, tape drive mechanism, and other appropriate hardware. .

  FIG. 4 is a diagram showing a software configuration of the tape file management system 100 according to the embodiment of the present invention. The computer system 110 shown in FIG. 4 includes a tape file system 200 as a software configuration. Note that the configuration shown in FIG. 4 corresponds to a mode in which the tape library 130 shown in FIG. 2B is used. A similar software configuration can also be achieved by using the tape drive 132 alone shown in FIG.

  The tape file system 200 enables the tape library 130 to be mounted on the tape medium 134 as one file system. The tape medium 134 in the tape library 130 is typically accessed as a subdirectory in the tape library. The tape file system 200 operates on the computer system 110 connected to the tape drive 132, and accesses a recording area on the tape medium 134 via an LTFS file system interface.

  In the mode using a single tape drive shown in FIG. 2A, the tape file system 200 is mounted on the tape medium 134 as one file system. Similarly, the tape file system 200 runs on the computer system 110 connected to the tape drive 132 and provides access to a recording area on the tape medium 134 via the LTFS file system interface.

  The tape file system 200, as a module for managing the writing and reading of meta information according to the embodiment of the present invention, receives a meta information writing request 210 and processes it, and reads the meta information. And a meta information reading unit 220 that receives and processes the request. The write request is generated, for example, when predetermined meta information is newly created in data such as a predetermined file, directory, or symbolic link from an application, or when predetermined meta information of predetermined data is updated. A read request is generated when, for example, predetermined meta information of predetermined data is referred from an application.

  The tape file system 200 can further include an index management unit 230. The index management unit 230 is a module that maintains and manages the index of the mounted tape medium 134 on the memory 12. When the tape medium 134 is mounted in the tape library 130, the index management unit 230 reads the latest index file from the mounted tape medium 134 and expands the index data 202 on the memory 12 as necessary. The index management unit 230 further updates the latest index file based on the index data 202 on the memory 12 when the tape medium 134 is unmounted in the tape library 130, when the file is closed, or periodically. To the mounted tape medium 134.

  More specifically, the meta information writing unit 210 includes a storage location determination unit 212, an index storage processing unit 214, and an external writing processing unit 216.

  In response to a write request for predetermined meta information of predetermined data (for example, a predetermined file; the same applies hereinafter), the storage destination determination unit 212 determines that the capacity of the meta information related to the write request is stored outside the index. It is determined whether or not a standard (hereinafter sometimes referred to as an external storage standard) is satisfied. The external storage standard can be defined as, for example, a threshold (for example, 4 KB or more) for the capacity of meta information.

  When it is determined that the external storage standard is satisfied, the external write processing unit 216 writes the main data of the meta information related to the write request in the storage area on the DP of the tape medium 134 that is outside the index. When it is determined that the external storage standard is satisfied, the index storage processing unit 214 uses the logical location information indicating the storage area on the DP in which the main data of the meta information is written as the predetermined meta information of the predetermined data. Store in the index. At this time, the logical position information is temporarily stored in the index data 202 on the memory 12, and the latest index file is written on the tape medium 134 at an appropriate time. Further, when it is determined that the external storage standard is not satisfied, the index storage processing unit 214 stores the main data of the meta information in the index as the predetermined meta information of the predetermined data.

  When the meta information writing unit 210 completes the writing of the meta information related to the write request, the meta information writing unit 210 notifies the request source, for example, the application, of the completion of the writing of the meta information. The storage destination determination unit 212, the index storage processing unit 214, and the external write processing unit 216 constitute a storage destination determination unit, a storage processing unit, and a write processing unit in the present embodiment, respectively.

  More specifically, the meta information reading unit 220 includes a storage source determination unit 222, an index acquisition processing unit 224, and an external reading processing unit 226.

  In response to a read request for predetermined meta information of predetermined data (for example, a predetermined file, the same applies hereinafter), the storage source determination unit 222 stores the meta information related to the read request outside the index. It is determined whether or not. When the index does not include the main data of the meta information related to the read request and the logical position information is stored, it is determined that the index is stored outside the index.

  When it is determined that the index acquisition processing unit 224 is not stored outside the index, the index acquisition processing unit 224 acquires the main data of the meta information related to the read request from the index. Further, when it is determined that the index acquisition processing unit 224 is stored outside the index, the index acquisition processing unit 224 acquires logical position information indicating the storage area on the DP outside the index from the index. At this time, logical position information is acquired as predetermined meta information of predetermined data from the index data 202 developed on the memory 12 when the tape medium 134 is mounted. When it is determined that the external read processing unit 226 is stored outside the index, the external read processing unit 226 reads the main data of the meta information related to the read request from the storage area specified based on the acquired logical position information. .

  When the meta information main body data of the meta information related to the read request is obtained, the meta information reading unit 220 passes the obtained meta information to a request source, for example, an application. The storage source determination unit 222, the index acquisition processing unit 224, and the external read processing unit 226 constitute a storage source determination unit, an acquisition processing unit, and a read processing unit in the present embodiment, respectively.

  In the above description, when storing outside the index, only the logical position information of the meta information main body data and the logical position information is held in the index, and all the main body data of the meta information related to the write request is stored. Is written in the storage area on the DP outside the index. However, the present invention is not limited to such an embodiment. In another embodiment, a part of the main body data of the meta information related to the write request is held in the index, and the remaining part of the main body data of the meta information is stored. May be written in the storage area on the DP outside the index. In this case, in response to the meta information read request, a part of the main data of the meta information related to the read request is acquired from the index, and the remaining part of the main data of the meta information is stored in the storage area on the DP outside the index. It becomes the aspect which reads and responds from.

  Details of the meta information writing method according to the embodiment of the present invention will be described below with reference to the flowchart shown in FIG. 5 and the data structure shown in FIG. FIG. 5 is a flowchart showing a meta information writing method executed by the computer system according to the embodiment of the present invention.

  The process shown in FIG. 5 is started from step S100 in response to a request for writing predetermined meta information from the application to the tape file system 200, for example. In step S <b> 101, the computer system 110 uses the storage destination determination unit 212 to determine whether or not the capacity of the meta information related to the write request satisfies the external storage standard. For example, when the meta information related to the write request has an arbitrary length such as a fork, and the capacity of the content of the meta information exceeds a predetermined threshold, it is determined that the external storage standard is satisfied.

  If it is determined in step S101 that the external storage standard is not satisfied (NO), the process branches to step S102. In step S102, the computer system 110 writes the meta information related to the write request to the index (index data 202 on the memory 12) by the index storage processing unit 214 as usual, and ends this processing in step S105.

  On the other hand, if it is determined in step S101 that the external storage standard is satisfied (YES), the process branches to step S103. In step S103, the computer system 110 uses the external write processing unit 216 to write the main data of the meta information to the DP of the tape medium 134 that is outside the index, and obtains the logical position information thereof, as with the file main data. To do.

  More specifically, in step S103, the computer system 110 instructs the tape drive 132 to write the main data of the meta information on the DP of the tape medium 134. Then, as a response to the command, the computer system 110 acquires, from the tape drive 132, logical position information indicating the storage area in which the main data of the meta information is written.

  In step S104, the computer system 110 uses the index storage processing unit 214 to index the logical location information indicating the storage area in which the main data of the meta information related to the write request is written (index data 202 on the memory 12). In step S105, the process is terminated.

  FIG. 6 is a diagram illustrating a data structure of an index file created based on the meta information writing method shown in FIG. FIG. 6 illustrates an XML format index file that is generated based on the index data 202 on the memory 12 and stored in the IP of the tape medium 134.

  As shown in FIG. 6, each index includes one or more pieces of meta information sandwiched between a start tag (for example, <name> tag) and an end tag (for example, </ name> tag). In FIG. 6, a portion surrounded by a broken line indicates a location where logical position information indicating main data of meta information is stored for meta information that satisfies the external storage criteria. In the example shown in FIG. 6, meta information including logical location information for each of one or more meta information satisfying the external storage standard is stored in a portion sandwiched between tags <out_of_index_metadata> </ out_of_index_metadata>.

  As the logical position information, in the example shown in FIG. 6, key information sandwiched between tags <key> </ key>, partition information sandwiched between tags <partition> </ partition>, tags <startblock> <// It includes start block information sandwiched between startblock>, offset information sandwiched between tags <byteoffset> </ byteoffset>, and count information sandwiched between tags <bytecount> </ bytecount>. When the metadata is not divided, the count information indicates the data size. Further, the metadata may be stored separately, and when divided, a plurality of start block information, offset information, and count information are written, and the count information is divided from the offset of the start block. This indicates how many bytes of the corresponding data are written.

  For other meta information (for example, a file attribute such as a length sandwiched between <length> </ length> tags), the meta information main body is stored instead of the logical position information.

  As shown in FIG. 6, in the index file created based on the meta information writing method according to the embodiment of the present invention, the main data of meta information having a capacity satisfying the external storage standard is written on the DP. Logical position information indicating the storage position is held.

  Details of the meta information reading method according to the embodiment of the present invention will be described below with reference to the flowchart shown in FIG. 7 and the data structure shown in FIG. FIG. 7 is a flowchart showing a meta information reading method executed by the computer system according to the embodiment of the present invention.

  The process shown in FIG. 7 is started from step S200 in response to a request for reading predetermined meta information from the application to the tape file system 200, for example. In step S <b> 201, the computer system 110 determines whether or not the meta information related to the read request is externally stored by the storage source determination unit 222.

  For example, in the case of meta information such as a file name sandwiched between tags <name> and </ name> shown in FIG. 6, the meta information main body is included in the meta information, and therefore it is not stored externally. Determined. On the other hand, meta information in a portion sandwiched between tags <out_of_index_metadata> </ out_of_index_metadata> (for example, meta information sandwiched between tags <metadata1> </ metadata1>) has a logical position in Since information is included, it is determined that the information is stored externally.

  If it is determined in step S201 that the corresponding meta information is not the logical location information but the main body of the meta information and is not stored externally (NO), the process branches to step S202. Is done. In step S202, the computer system 110 reads the meta information related to the read request from the index (index data 202 on the memory 12) by the index acquisition processing unit 224 as usual, and ends the processing in step S205.

  On the other hand, if it is determined in step S201 that the logical location information, not the main body of the meta information, is held as the corresponding meta information and is stored externally (YES), step S203 is performed. Processing is branched to. In step S203, the computer system 110 uses the index acquisition processing unit 224 to obtain logical position information indicating the storage area in which the main data of the meta information related to the read request is written from the index (index data 202 on the memory 12). read out.

  In step S204, the computer system 110 causes the external read processing unit 226 to perform the meta data relating to the read request from the storage area indicated by the DP logical position information of the tape medium 134 that is out of the index, similar to the file main body data. The main data of the information is read out, and this process is terminated in step S205.

  More specifically, in step S204, the computer system 110 gives the logical position information on the DP on the tape medium 134 to the tape drive 132 to instruct reading. Then, as a response to the command, the computer system 110 acquires the main data of the meta information written in the storage area specified by the logical position information from the tape drive 132 as a read result.

  In the above description, the writing and reading of the meta information is performed on the index data 202 on the memory 12. Hereinafter, with reference to FIG. 8, processing for storing the index data 202 including the logical position information on the memory 12 on the tape medium 134, and an index file including the logical position information on the tape medium 134 are stored on the memory 12. The processing to be developed will be described.

  FIG. 8 is a diagram illustrating a configuration in which the index management unit 230 according to the embodiment of the present invention maintains index data on the memory 12. As shown in FIG. 8, when the tape medium 134 is mounted, the computer system 110 reads the latest version of the index file from the IP on the tape medium 134 and stores the index data 202 on the memory 12 included in the computer system 110. expand. Then, when a synchronization operation such as unmounting of the tape medium is performed, the computer system 110 overwrites the IP on the tape medium with the latest version of the index based on the index data 202 of the memory 12.

  When a write request for meta information of predetermined data is generated, the meta information main body on the DP of the tape medium 134 is not written in the index on the IP of the tape medium 134 but first in the index data 202 on the memory 12. The logical position information is written. Further, when a read request for the meta information of predetermined data is generated, the index on the IP of the tape medium 134 is not read, and the meta information main body on the DP of the tape medium 134 based on the index data 202 on the memory 12. Is accessed.

  As described above, the logical position information is written into the index data 202 on the memory 12 in response to the meta information write request. The index data 202 including the logical position information is stored at the timing of the synchronization operation. Thus, it is stored in the tape medium 134 as an index file. Further, as described above, the logical position information is read from the index data 202 on the memory 12 in response to the meta information read request. The index data 202 including the logical position information is stored in the tape medium 134. Is mounted on the memory 12 by reading the index file on the tape medium 134 at the timing of mounting.

  FIG. 9 is a diagram for explaining the effect of the meta information storage method provided by the tape file management system 100 according to the embodiment of the present invention, together with the data structure. As shown in FIG. 9, the volume layout configured according to the present embodiment is similar to the prior art shown in FIG. 1, and includes an IP 310 that stores an index that holds one or more pieces of meta information relating to each of one or more pieces of data. It includes a DP 320 that stores main data of each of one or more data and additionally stores index history 322-1 and 322-2. However, as is clear from comparison with FIG. 1, the DP 320 further stores meta information main body data 324 having a capacity satisfying a predetermined standard. As described above, the latest indexes 312 and 322-2 written on the tape medium 134 include the logic indicating the storage location on the DP 320 where the main body data of the meta information having the capacity satisfying the predetermined standard is written. Location information is retained.

  By adopting the meta information storage method, the problem that the index capacity becomes too large and the index file does not fit in the IP and the problem that it takes time to write the index file to the IP are prevented. Further, when the index history is accumulated in the DP, the problem that the DP capacity is consumed in large quantities by the meta information and the DP resources are compressed is prevented. Furthermore, since only the logical position information is maintained for the large amount of meta information in the index data 202 on the memory 12, the problem that the index squeezes memory resources can be prevented. The meta information storage method described above is particularly useful when it is desired to store meta information such as a fork on a tape medium that can have a large capacity.

  Here, it is assumed that 1000 files having a file size of 1 GB are written on the LTO5 tape medium while performing a synchronization operation (Sync Operation) once every five minutes. In this case, considering the writing speed of LTP5, it is estimated that a total of 35 synchronization operations will be performed before writing 1000 1 GB files. At this time, in the first synchronization operation, for example, 29 files of meta information are included in the index, while in the second synchronization operation, for example, 29 additional meta information of 29 files is added. The index size progressively increases so that is included in the index. In consideration of such a progressively increasing index size, an average of 500 files of meta information is stored in the index.

  If each file has a fork of 20 MB, when this fork is written in the index, the capacity of the fork in the index is 10 GB on average (20 MB × 500 files on average). If it is written 35 times, 350 GB is written. It will consume capacity.

  In contrast, by applying the meta information storage method according to the embodiment of the present invention described above, the capacity of the fork is 20 GB (= 20 MB × 1000) outside the index because the fork body is stored outside the index. ) And the logical position information data in the index (the logical position information itself has a small capacity and is a small capacity even if written 35 times). In other words, 300 or more 1 GB files can be written. Further calculation of the required update time of the index file shows that when the transfer rate is 100 MB / sec, the update time is reduced by about 1 hour (= 3300 [sec] = 330 [GB] / 100 [MB / sec]). Will be able to.

  As described above, according to the embodiment of the present invention, in a storage medium in which an index that holds meta information of each of one or more pieces of data is stored, the enlargement of the index is prevented and the resource is not compressed. A meta information writing method, meta information reading method, file management system, computer system, program, and data structure capable of storing information can be provided.

  In the above-described embodiment, the tape medium is described as a suitable example as the storage medium. The meta-information storage method according to the embodiment of the present invention is particularly effective for a tape medium. However, a medium other than a tape medium is not excluded as a storage medium. In other embodiments, a sequential access is performed. The present invention can be applied to any kind of storage medium that is a medium.

  The above-described functions of the present invention can be realized by a device-executable program written in a legacy programming language such as an assembler, C, C ++, C #, Java (registered trademark) or an object-oriented programming language, and is readable by a device. It can be stored in a recording medium and distributed or transmitted.

  Although the present invention has been described with specific embodiments, the present invention is not limited to the embodiments, and other embodiments, additions, changes, deletions, and the like can be conceived by those skilled in the art. It can be changed within the range, and any embodiment is included in the scope of the present invention as long as the effects and effects of the present invention are exhibited.

DESCRIPTION OF SYMBOLS 10 ... CPU, 12 ... Memory, 14 ... Storage device, 16 ... Network interface device, 18 ... Input / output I / F, 100 ... File management system for tape, 110 ... Computer system, 130 ... Tape library, 132 ... Tape Drive 134 134 Tape medium 200 Tape file system 202 Index data 210 Meta information writing unit 212 Storage destination determination unit 214 Index storage processing unit 216 External writing processing unit 220 ... meta information reading unit, 222 ... storage source determination unit, 224 ... index acquisition processing unit, 226 ... external reading processing unit, 230 ... index management unit, 310 ... IP, 312 ... index, 320 ... DP, 322 ... index, 324 ... meta information

Claims (19)

A meta information writing method for writing meta information to a storage medium in which an index that holds meta information of each of one or more pieces of data is stored, the computer system comprising:
In response to a write request for meta information, determining whether the capacity of the meta information related to the write request satisfies a criterion for non-index storage;
Writing at least a part of the main body of the meta-information related to the write request to the storage area outside the index when it is determined that the criterion for storage outside the index is satisfied;
Storing in the index position information indicating the storage area in which at least a part of the main body is written. In the meta information writing method, the computer system
The meta information writing method according to claim 1, further comprising: storing the main body of meta information related to the write request in the index when it is determined that the non-index storage criterion is not satisfied. The index is maintained on a storage device included in the computer system, and the meta information writing method is performed by the computer system.
The meta information writing method according to claim 1, further comprising: overwriting the latest version of the index in an index section partitioned on the storage medium. The step of writing to the storage area outside the index includes:
Instructing a storage medium driving device to write at least a part of the main body to a data section partitioned separately from the index section on the storage medium;
The meta information writing method according to claim 3, further comprising: acquiring position information indicating the storage area on the storage medium in which at least a part of the main body is written from the storage medium driving device.   5. The meta information writing method according to claim 4, wherein a main part of each of the one or more data and a history of the index are additionally stored in the data portion of the storage medium.   The meta information writing method according to claim 1, wherein the storage medium is a sequential access medium.   The meta information writing method according to claim 1, wherein the storage medium is a tape medium. A meta information reading method for reading meta information from a storage medium in which an index for holding meta information of each of one or more data is stored, wherein the computer system includes:
In response to a read request for meta information, determining whether the meta information related to the read request is stored outside the index;
Obtaining a position information indicating a storage area outside the index from the index when it is determined that the data is stored outside the index; and
Reading at least part of the main body of the meta information related to the read request from the storage area specified based on the acquired position information. In the meta information reading method, the computer system
The meta information reading method according to claim 8, further comprising a step of acquiring, from the index, a main body of meta information related to the reading request when it is determined that the information is not stored outside the index. In the meta information reading method, the computer system
The meta information reading method according to claim 8, further comprising: reading the latest version of the index from an index section partitioned on the storage medium and maintaining it on a storage device included in the computer system. The step of reading at least a part of the main body of the meta information relating to the read request includes:
Instructing the storage medium driving device to provide position information indicating the storage area on the data section partitioned from the index section on the storage medium and instructing reading;
The meta information reading method according to claim 10, further comprising: acquiring a reading result of at least a part of the main body from the storage medium driving device.   The meta information stored outside the index is written in the data portion of the storage medium so that the capacity of the meta information satisfies the criteria for storage outside the index, and the data portion of the storage medium further includes: The meta information reading method according to any one of claims 8 to 11, wherein a main body of each of the one or more data and the history of the index are additionally stored. A file management system that performs file management of a storage medium in which an index that holds meta information of each of one or more pieces of data is stored on a computer system,
In response to a request for writing meta information, a storage destination determination unit that determines whether the capacity of the meta information related to the write request satisfies a criterion for non-index storage;
A write processing unit for writing at least a part of the main body of the meta-information related to the write request to the storage area outside the index when it is determined that the criterion for storage outside the index is satisfied;
A storage processing unit that stores position information indicating the storage area in which at least a part of the main body is written in the index. A file management system that performs file management of a storage medium in which an index that holds meta information of each of one or more pieces of data is stored on a computer system,
In response to a read request for meta information, a storage source determination unit that determines whether the meta information related to the read request is stored outside the index;
An acquisition processing unit for acquiring, from the index, position information indicating a storage area outside the index when it is determined that the data is stored outside the index;
A read processing unit that reads at least a part of the main body of the meta information relating to the read request from the storage area specified based on the acquired position information.   The file management system according to claim 13 or 14, wherein the file management system includes one or more storage medium driving devices, and the computer system includes one or more computers. A processor;
A memory in communication with the processor;
An interface device for connecting to a storage medium driving device that drives a storage medium in which an index that holds meta information of each of the one or more data is stored, wherein the processor comprises:
In response to a write request for meta information, it is determined whether the capacity of the meta information related to the write request satisfies a criterion for storage outside the index,
When it is determined that the non-index storage criterion is satisfied, at least a part of the meta information body related to the write request is written to the non-index storage area;
A computer system configured to store, in the index, position information indicating the storage area in which at least a part of the main body is written. A processor;
A memory in communication with the processor;
An interface device for connecting to a storage medium driving device that drives a storage medium in which an index that holds meta information of each of the one or more data is stored, wherein the processor comprises:
In response to the read request for meta information, it is determined whether the meta information related to the read request is stored outside the index;
When it is determined that the data is stored outside the index, position information indicating a storage area outside the index is acquired from the index;
A computer system configured to read at least a part of a main body of meta information relating to the read request from a storage area specified based on the acquired position information.   15. A program for realizing the file management system according to claim 13 or 14, wherein the program causes a computer system to function as each unit according to claim 13 or 14. An index partition that stores an index in which one or more meta information about each of the one or more files is held;
A data partition for storing body data of each of the one or more files and additionally storing the history of the index;
The data partition further stores main data of meta information having a capacity satisfying a predetermined standard, and the index is the data partition in which main data of meta information having a capacity satisfying the predetermined standard is written. A data structure characterized by holding position information indicating an upper storage position.